INTEGRATION OF CORPUS TECHNOLOGIES INTO TEACHING A PRACTICAL RUSSIAN TRANSLATION COURSE FOR FOREIGN LEARNERS

Abstract

In the context of globalization, the demand is growing for translators capable of producing natural and stylistically accurate texts in Russian as a foreign language. However, traditional methods based on dictionaries do not develop students’ skills in analyzing context and collocations, leading to unnatural translations. The relevance of this research stems from the need to overcome this discrepancy by integrating corpus technologies, which provide access to authentic texts for developing a “feel for the language.” The objective is to justify and experimentally test the effectiveness of a translation teaching method using the Russian National Corpus (RNC) and Sketch Engine, comparing it with the traditional approach according to the criteria of lexical-semantic accuracy, collocational adequacy, speed, and student confidence. The experiment involved 64 foreign students at the B2-C1 level, divided into a control group (CG) and an experimental group (EG) of 32 people each. The initial level was verified by a preliminary test translating a text (1,500 characters). The CG was taught traditionally (dictionaries, consultations), while the EG actively used corpora to analyze collocations and semantics. Materials included journalistic and business texts. Evaluation was carried out using unified criteria; the data was processed using mathematical statistics methods (mean values, standard deviation, percentage increase). The results showed a significant advantage for the EG: lexical-semantic accuracy increased by 31.99% (versus 11.14% in the CG), collocation errors decreased by 72.19% (versus 29.89%). Translation speed in the EG was lower (291.3 words/hour versus 315.6), but confidence was higher (8.9 out of 10 versus 6.2). Within the EG, a correlation between corpus usage frequency and progress was confirmed: high activity resulted in a 45.81% increase. The discussion of results confirms that corpus technologies develop research skills, improving translation quality through empirical analysis. This shifts the focus from memorization to a contextual approach, reducing errors and increasing confidence. The methodology is recommended for integration into programs, with the development of assignments for regular corpus use. The conclusions emphasize the potential for modernizing teaching, but further research is needed on automating tools to speed up the process.